A Study on Peer Data Management Systems for P2P Networks

碩士 === 國立中興大學 === 資訊管理學系所 === 97 === In recent years, peer-to-peer (P2P) computing has attracted much attention in both research and industrial communities. Many P2P systems have been proposed such as CAN, Pastry, Chord, Tapstry, etc., but they can only support exact-match lookups. There is a new ar...

Full description

Bibliographic Details
Main Authors: Yao-Tsan Chang, 張耀燦
Other Authors: Eric Jui-Lin Lu
Format: Others
Language:en_US
Online Access:http://ndltd.ncl.edu.tw/handle/21993186986175053110
Description
Summary:碩士 === 國立中興大學 === 資訊管理學系所 === 97 === In recent years, peer-to-peer (P2P) computing has attracted much attention in both research and industrial communities. Many P2P systems have been proposed such as CAN, Pastry, Chord, Tapstry, etc., but they can only support exact-match lookups. There is a new arena of P2P research, called peer data management system (PDMS), they use metadata which provide great flexibility for annotating resources to describe the shared resources and can support complex queries. However, most of them use the keywords to locate the corresponding resources, and only a few of them can support RDF queries to retrieve the RDF documents that describe the requested resources. In this thesis, we propose a distributed RDF repository called RDF-Chord. In RDF-Chord, nodes are assigned into the appropriate ring sets based on the shared resources and a query is forward only to a subset nodes that stores the indexes of the requested resources by using the appropriate ring sets. Via detailed simulations, we show that our approach is highly scalable and can resolve RDF queries efficiently.